© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 41219
    Prev
    Next
    INDEX
    Explanations

    variations of the phrase "make up."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æijĺè¦ģ
    -0.16
    opper
    -0.16
    tsky
    -0.15
    زا
    -0.15
    t
    -0.15
    tet
    -0.14
    cko
    -0.14
    uzzi
    -0.14
    asco
    -0.14
    adium
    -0.14
    POSITIVE LOGITS
     for
    0.25
    ping
    0.21
     ground
    0.20
    pers
    0.20
    ped
    0.19
     lost
    0.19
    /down
    0.17
    rightness
    0.17
    datable
    0.17
    ToBounds
    0.17
    Activations Density 0.014%

    No Known Activations